His ideas were few but enormous: a definition of computation itself, a proof of its limits, and a question that still has no settled answer — can a machine think? This book is what he actually argued, and the line everyone half-remembers.
An abstract head-and-tape that defines what "computable" means.
computationOne machine that can simulate any other — the stored-program computer.
one for allThe halting problem: some questions no machine can ever answer.
the boundaryJudge a machine's mind by whether it can pass for human in conversation.
the testA head that reads a symbol, writes a symbol, and moves along an endless tape by simple rules.
from "On Computable Numbers" (1936)
so "computable" got a precise, mechanical definition.
+1 the tape is imagined as infinite — the model is about what's possible in principle, not buildable hardware.
A single Turing machine that, fed a description of any other, behaves exactly like it.
idea a machine that reads programs
so one device can do every computable task — the computer.
+1 this is the stored-program idea: the program is just data the machine reads, like any other input.
Anything a human can compute by rote, a Turing machine can compute too.
with Alonzo Church, independently
so the Turing machine captures computation itself.
+1 it's a thesis, not a theorem — accepted because nothing has ever broken it, not because it's proven.
No machine can decide, in general, whether an arbitrary program will ever stop.
result undecidability
so he proved hard limits on what computers can ever do.
+1 this was the original point of the 1936 paper — the computer fell out of a proof about impossibility.
If a machine can converse so a judge can't tell it from a human, he argued, call it intelligent.
from "Computing Machinery and Intelligence" (1950)
so he reframed "can it think?" as something testable.
+1 he sidestepped defining "think" — replacing a vague question with a concrete behavioral one.
He took up Ada's claim that machines can only do what they're told — and pushed back.
he named it "Lady Lovelace's Objection"
so the two pioneers debate across a century, through him.
+1 he argued machines could surprise us and learn — a reply still echoing in today's AI.
He suggested building a "child machine" and teaching it, rather than programming it fully.
idea machines that learn from experience
so he anticipated machine learning decades early.
+1 "instead of producing a program to simulate the adult mind, why not… simulate the child's?" — strikingly modern.
His last work asked how chemistry alone can produce the patterns of living things.
paper on morphogenesis (1952)
so he founded a strand of mathematical biology.
+1 his reaction-diffusion equations still help explain how animals get their spots and stripes.
enihundua series · book no. 1 · he defined the computer, then asked if it could think · the question is still open